For our latest webinar, we spoke to Ross Tavendale of Type A Media, about data journalism and how to use simple datasets to make stories that get links. Watch the recorded webinar below and read on to get a practical guide to off page SEO.
It can often be quite challenging to create interesting stories that will get shared and linked to. Sometimes you need to manufacture the news.
Data Journalism: How to Use Simple Datasets to Make Stories that Make You Famous
First, let’s talk about flop risk. Sometimes marketing campaigns work, sometimes they don’t. When they flop, that’s bad so we need to reduce our risk as much as possible. So, a campaign that’s got a high flop risk would be pitching to the BBC or the New York Times, for example, to get a piece in the newspaper. However, these pitches would also have a low algorithmic risk.
“I’ve never had something with the BBC get me a penalty and I don’t think it ever will.”
However, if you’re consistently guest posting and creating a footprint for yourself, you’re more likely to get those links, but it’s going to create more algorithmic risk.
What we’re going to be talking about in this post about data journalism, is at the far end of this graph, so high flop risk, zero algorithmic risk. In order to reduce the flop risk we need to come up with a framework so that our campaign is successful regardless of what actually happens when we pitch.
“We are an agency so we reserve the right to have acronyms for everything we do.”
This is the methodology they use at Type A Media.
D.I.S.C.O. = Discovery & Planning, Ideation, Sign Off, Creation, Outreach.
It’s important to note that sign off is bang in the middle. So before the client actually sees anything, if you’re holding an ideation session, make sure the client or your boss, or whoever gets to sign off on your campaign, is not actually in the room. They should have absolutely no say in what goes into an ideation session as they have too much power and that makes it harder for you to come up with good ideas.
Validate & Verify
Before you take any ideas to get signed off, there are a few things you need to do to make the process a lot smoother.
First, you should cold pitch your ideas to journalists. Your ideas are based on your subjective opinion but if you get journalists interested, then you have something to back your ideas up. Instead of asking your client to pick their favourite idea, tell them which one had the best reaction from journalists.
Make sure there’s enough data available for the story before you pitch. You may have a good idea, but if there’s no data or limited data for the story, then it’s not going to work, or it’s going to take a lot longer and may not be worth the time.
Confidence in the data is also very important. First-party data supplied by your client is the best. If you’re going to need to scrape a whole bunch of websites to find the data, then confidence in that data goes down and down as it’s harder to wrangle it in order to get a cohesive dataset.
It may be a winner when it comes to PR, but a nightmare to actually put together. It may take 2 months to make the asset if there’s not a lot of data readily available and by that point your time for PR may have run out.
Validate & Verify
This is a simple table to quickly work out what data journalism will work and what won’t. We use something similar at In Marketing We Trust.
Data Journalism RAP Sheet
Next, you’ll want to create a Data Journalism RAP (Research, Angles, Pitch) sheet. It should look something like this:
Many people reaching out to journalists will start writing what they think they want to hear. This is far too late in the process to be finding your headlines. You need to know what they are well before you start pitching journalists.
Working on a RAP sheet will not only help you put together your ideas, it will also help you to get a quicker sign-off.
This is a really simple way to document your research, angles and pitches in one place. Sometimes when you write down your headline and read it back to yourself it’s not so great. If you’ve got dedicated time to come up with a better headline, it’s going to be a better campaign, so don’t leave it too late in the process.
“Get to that point too late and you’re being reactive.”
3 Angles, 3 Headlines Gets You 9 Pitches
Because your story is data lead, you can cut it in multiple ways. If you come up with 3 different angles and 3 headlines, that’s 9 completely different pitches.
You’ll want a national angle, a local angle, a trade angle with different headlines. This way, if the journalist you pitch the national headline to doesn’t like it, you have 8 different groups of journalists you can pitch to with completely different angles so your flop risk is dramatically reduced.
That means you only need to win 11% of the time.
Have you ever seen an ad on TV you thought was rubbish? Of course you have. That ad had to go through creative direction, art direction, the whole deal and then someone spent millions of dollars getting it on TV. But it still sucks. The thing is, it doesn’t matter. The risk is removed by the amount of people who are seeing the ad. If you show an add to millions of people and 90% think it’s crap, it doesn’t matter, a hundred thousand people like it. Ads like this reduce their flop rate by showing it to so many people.
The way we reduce our flop risk is by doing the same thing only in this case, we’re showing it to lots of journalists, rather than getting media impressions.
Data Journalism: Creating Your Content
Now, let’s dive into actually creating your content.
Start with the Audience
First, start with the audience. It’s very important to work out who you’re actually talking to. The best place to go is straight into Facebook Audience Insights. You’re not going to be looking for women who are between 25 and 34 who live in Brisbane who own a lawn mower. We don’t care about that. What you care about is weird outlier facts about the person.
You’re going to jump on these outlier facts and make an interesting story about them.
“I don’t care about the core demographic or the usual stuff, I wanna see something weird I can make a cool story about.”
Example: Casino games
When we put into Facebook Audience Insights “casino games” the audience that actually consumes them are 93% more likely to be engaged and 58% less likely to be married.
Story idea: A piece on the size of wedding rings.
“The client thought we were absolutely mental but it ended up earning 40-50 links because it spoke to the audience.”
Example: Investment banking
Investment bankers love money, of course, but they also love Nandos. They also like to smell and dress like a teenager, so they like Lynx deodorant and glam sportswear. They also like pages like “What James Bond character are you?”.
So next, we write assumptive statements on a whiteboard so when we start coming up with ideas we have something to test the assumption off of, and if it fails the assumption, we don’t use the idea.
This way, we’re testing against the person and data, not our random subjective opinions.
Data Journalism Ideation Framework
Just like a pop song has a formula, so does content.
Content that covers identity is something that expresses your identity better than you can.
An example of this type of content would be “It’s Hard Out There for a Lefty”. Being left-handed is an identity factor so when you create content like that you’re naturally going to get a lot of shares from people who identify. This type of content works really well for clients with no data to share so we just do original research and package it up as news.
Utility is much more common for building out content for links. Think calculators, so mortgage calculators and other tools for example. To make it interesting, you can simply change the way the content looks, a data journalism example using the utility framework could be “What 2000 Calories Looks Like in Fast Foods”.
This is just a data table but we’re taking it and presenting it in a completely different way that’s useful and interesting and it’s much easier to get it live in the press.
Another example was for a skincare client that had no useful data so it was summer coming up so we got a dermatologist to rank booze from best to worst for your skin (spoiler alert: all booze is bad).
This last one is really difficult to do. Social information takes an idea that you already believe in and it breaks it or confirms it. An example of this would be “Londoners Found to be Most Miserable People in the UK”. When we’re pitching to people in the North of England who typically don’t like people from the big city, it does really well in Northern English papers. So the concept is, we’re taking something you already believe in and saying well actually that’s true, or that’s completely false.
It doesn’t have to require large datasets, the data can actually come from something as basic as google trends.
Data Journalism Example Using Google Trends
Now if you look at this in Google Trends and zoom out technically that headline would be Allergies on the Rise Globally, because every year there’s an uptick. Now, when you dig into the data, it’s young people who are searching for ‘allergies’ more. So if we want to create an information piece to pitch to right wing media in the UK, we can run it through the framework and we know that the Daily Mail for example will like social information pieces that confirm millennials are snowflakes.
The reason this methodology works is because it works in reverse of how you usually get content discovered, so the content finds the audience, the audience don’t find the content. What we mean by that is because you’re a lefty, because you have ginger hair, because you think millennials are snowflakes, that will automatically get in your filter bubble of news so that content automatically finds you.
Data Journalism Ideation framework
Here’s an example of some of the things we’d typically put under each of these categories. The social information pieces are not as brand safe as the rest, so just be very careful and get brand managers to sign off before it goes live.
Drill Down-ability for Multiple Headlines
We’re coming up with this stuff to reduce flop risk, so we need to come up with a really big concept and drill it all the way down into something really small to get multiple bites. So not only do we have 9 different pitches from national to local to trade inside each of those verticals we’ve got big publications all the way down to small.
Take gaming for example, Type A Media ran a piece for a client where they looked at the historic value of computer games. There’s a piece of news that says Mario Kart is worth 30k so we took an average amount of computer games that people were sitting on and selling on eBay and worked out how much ‘money’ they were sitting on as an asset with their old computer games.
So in this example, we go broad with gamers, then slightly more niche with talking to computer game magazines, then go into community stuff, so blogs and really small magazines, and then single-person-operated websites with content for fanatics. Obviously there’s less of them but that means you don’t need to get as personal with the pitching. So what we’re trying to do here is reduce flop risk by broadening the scope as much as humanly possible.
3 Types of Data Sources That Can Be Used To Tell a Story
Private Data from Analytics & CRM
The best type of data source by far is private data from your analytics or CRM. If you have a BI team that can provide you with actual industry information that is absolute gold. Include a quote in there and you’ll get links every time.
It doesn’t have to be really interesting either, so for a client of Type A they ended up just taking their Google Analytics data and looked at where the most apprenticeships were up and down the country and that was enough to get them published.
Public and Government Data
We like government datasets as well, in fact this is probably the one we use the most. So for most countries, their government websites have these massive public datasets that are free to download. Now you need to have data scientists to actually use it because there are millions of lines and you actually often can’t physically open it in a csv because it’s too much to load but it’s all there and it’s free.
Because it’s a government source it’s also already pre-validated so when you pitch that to an editor you can say it’s government data and they’re sold.
Potential Data from Surveys & Pro Commentary
It’s really hard to pitch a survey to a national newspaper and for them to care because it’s usually not a big enough sample size. The questions are presented in a biased way to get the information we want for our piece, so it’s less interesting to people publishing the news but it’s still an option if you’re struggling.
What to do with Your Private Data
Mine Your Own Analytics
So for one of Type A’s clients, they went into their analytics and there’s this little place called East Killbride, an old mining town really damaged by the industry leaving. But this is the best place for young people to get jobs. This fits the ‘identity’ framework and it breaks ‘social information’ because it breaks your preconceived notions and that’s why it does so well.
Mashup Your Data with Other Sources
Mashup your data with other sources. Sometimes a client will give us something and it’s kind of ok but it’s not good enough to be newsworthy. So for example, Type A Media did something with a company who had an esports team. When looking at the data they quickly realised that the people playing FIFA esports were getting paid on average 75% more than professional football players in the premier league.
At the same time, the government had just released gender equality pay information in the UK, so they looked at all female football teams and found that 100% of people playing FIFA games get paid more than female football players. Then they found the head of the Women’s Football Association and pitched the story to her and it blew up.
How to Use Public Data
Make it Easier to Use
There was a point in the UK last year where Ryanair cancelled all their flights. How did they tell their customers? They put a PDF on their site and just essentially said “you deal with it – it’s on the PDF, go search for it. We don’t care.”
So, Type A took the PDF and created a tool where you could put in your flight number and have an automatic link to get a refund if your flight was on there. Then, they found a bunch of journalists that didn’t like Ryanair (spoiler alert: there’s loads of them), pitched to them and went viral.
Slice it and Serve it Differently
We’ve already mentioned the pay gap reports the UK government has started to put out. But what they do is just dump it in a massive database probably deliberately to stop people from reporting on it.
If you’re looking at a dataset you can barely open on your computer, it’s going to be very hard for someone to make a story about it so they just turned it into an interactive tool. So every time the government is publishing something they perhaps don’t want to be seen, we’ll take it and turn it into something and make it usable for journalists.
How to Find Public Data
Typically, we do site commands. So here for example ac.uk is a university website. So inurl:pdf will show all of the papers they’ve published and then add a keyword. You’d be surprised at the amount of publicly available documents that are from universities that can’t be found elsewhere, so definitely use this advanced search to find data for pieces.
Google Dataset Search
The obvious way to find public data is a Google dataset search which is really just powered by a couple of large database providers. It’s ok and it’s a good place to start off but it’s not the be all and end all.
Ross recommends using Statista and similar providers. There’s a lot of providers not covered in Google dataset search.
How to Use Potential Data
Freedom of Information Request
Most governments around the world will let you ask for information about a public body and the great thing about this is that when you’re pitching a journalist the story, you can just forward what the government sends.
For example, Type A Media contacted the ambulance service to see how many heart attacks occurred at football stadiums up and down the country, with the idea being who’s the most exciting football team as measured by the amount of people who have a heart attack in the stadium.
So they got the information from St. John Ambulance service, so straight from the government. They packaged it up and pitched it to the journalists and because the journalist could see the email thread coming from the government, they are much quicker to publish because they can clearly see the official source.
So Google Surveys are fine if you’re struggling and just doing local or trade news. But be warned, they cost a lot and they don’t usually provide a lot of data.
Data Journalism: A Practical Guide to Off Page SEO
You’ll never have a viral success with this data journalism approach but you will get repeatable data led campaigns. There’s going to be no surprises, you’re going to get a regular amount of links, you’re not going to blow up and go viral but you are going to get 30 to 50 good links every time. So it’s a relatively simple process when you know how.